Taylor 68A Signal Generator Waveform

[Nigel_H]

I have a Taylor 68A Signal Generator that is giving some rather odd output waveforms - see the photos, which are at the low ends of bands 1 to 4 (my scope hasn't enough bandwidth to examine the higher bands). The waveforms at the upper ends of the bands are similar, although there is a significant change in amplitude.

I looked at the various posts on this forum, and the first thing I noted is that the circuit of my model differs significantly from that in the manual which people have posted here: firstly, the modulator is an EC91 (rather than a double triode), secondly the HT recitifier is a 6X4 rather than a diode bridge, and thirdly there are a number of resistors associated with the oscillator coils. (The first two are irrelevant to this problem, of course, but the easiest to spot.)

As anyone who has one of these models will know, the oscillator assembly is a real ratsnest which makes it almost impossible to trace the circuit, but what I can see is that each of the lower-band coils has a separate series resistor of some 10s of kohms through which the HT is supplied, and another smaller resistor of some 10s of ohms, whose function is unclear (see photo). Given that they are carbon composition types, I wondered whether a shift in value might be the problem. The circuitry prevents me measuring the higher-valued resistors, but I measured the three low-valued ones; those on bands 1 and 2 measure correctly, while the 33 ohm resistor for band 3 measures as more like 22 ohms. The HT (measured at the choke feeding the oscillator) is about 130V, and all valves test good.

Does anyone have any suggestions, please, and/or a circuit diagram for this version? The serial number is 392020.

[HamishBoxer]

There is a circuit for the later 68A on radiomuseum.org. I have one of these and as yet untested.

[Nigel_H]

Thanks - unfortunately, that confirms that mine is the earlier model, for which there isn't a schematic there.

[chriswood1900]

This thread may help you.
https://www.vintage-radio.net/forum/...ad.php?t=89509

[HamishBoxer]

I wonder if the 67A is the same circuit or near enough?

[chriswood1900]

My guess 67a is likely to be similar certainly the valve line up is closer to the 67a than the later 68a that is available.

[Nigel_H]

Thanks for the help and comments. There's certainly some similarity with the 67A, but equally a lot of differences. I tried to trace the circuit (visually and with a multimeter), at least for one band, and it seems that the main differences from the 67A are that the unused coils are shorted out, and the output is taken from the secondary via a resistor (different for each band) rather than directly, and is also coupled via a 10n capacitor to the valve cathodes (both are connected together); I guess this must be some kind of level control using negative feedback? At any rate, having worked out the bit about the unused coil windings being shorted I was able to check the reststor values, and they're respectably close to what they should be. The tuning is pretty much spot on, so that rules out most of the capacitors.

Looking again at the waveforms, those on the oscillator grid and anode are reasonable sine waves, as is the one on the smaller coil winding. However, the oscillator cathode has a rather spiky waveform, and of course the output is a bit of a mixture, as seen before. I tried some Deoxit on the switch but that didn't help. I'll come back to this later and see if I can work out what's happening; maybe there's a dry joint somewhere, or somehow the unused section of the valve isn't disabled.

[chriswood1900]

I would not worry about the waveforms, most basic signal generators give fairly poor waveforms. If you look at outputs from the Nombrex and the TE20 they are all fairly horrible and in fact quite a few rely on the distortion to produce harmonics for the high frequencies. I'm not familiar with the Taylor models but would expect what you see to be typical.

[Freya]

This is my copy of the early 68am, my has serial number 180862

[factory]

Originally Posted by chriswood1900

I would not worry about the waveforms, most basic signal generators give fairly poor waveforms. If you look at outputs from the Nombrex and the TE20 they are all fairly horrible and in fact quite a few rely on the distortion to produce harmonics for the high frequencies. I'm not familiar with the Taylor models but would expect what you see to be typical.

The Taylor 67A I had a few years ago produced a better looking waveform the what the OP has from his 68A, the 67A was sold to another forum member a while back.


The only Taylor/Windsor signal generator I still have is the 65B, which is the first item of test gear I bought, it needs a full restoration.

David

[Nigel_H]

Thanks for the comments. I wouldn't have expected a wonderful waveform, but what seems really strange is that there's a perfectly decent sine wave on the resonant circuit, and it's just the coupling in of the valve cathode voltage (essentially, a large spike as the grid goes positive) that messes it up. It's interesting just how much the circuit changed between the models: in the 67A the resonant circuit is between anode and grid and the output from a separate winding, on the early 68am and later 68a the resonant circuit is between grid and ground (with an inductive tap for the cathode) with the output from a separate winding, whereas mine is like the 67A but with the output winding coupled to the cathode. I might see what happens if I disconnect the coupling capacitor!

While looking around my PC I came across the attached circuit, which exactly matches what I have and confirms the manufacturing date as around 1956. I'd forgotten I'd got that and have no idea how I found it, but the file date is 1/4/13 so I must have got it at a similar time to the signal generator itself.

[Nigel_H]

I'm pleased to say that there was a fault which I managed to find this morning and fix; I also found a copy of the manual wedged between two magazines(!), which I have scanned and attached here; I had previously annotated the schematic with most of the component values.

The problem was that I had misread the schematic before; there is capacitive coupling between the output and the valve cathodes, but the third switch wafer from the top on the schematic should have shorted the cathodes to ground for bands 1 to 3, disabling this. I found that the wire to the cathodes had come adrift; on resoldering it, I got a far better waveform!